An adaptive peer-sampling protocol for building networks of browsers

Abstract

Peer-sampling protocols constitute a fundamental mechanism for a number of large-scale distributed applications. The recent introduction of WebRTC facilitated the deployment of decentralized applications over a network of browsers. However, deploying existing peer-sampling protocols on top of We-bRTC raises issues about their lack of adaptiveness to sudden bursts of popularity over a network that does not manage addressing or routing. Spray is a novel random peer-sampling protocol that dynamically, quickly, and efficiently self-adapts to the network size. Our experiments show the flexibility of Spray and highlight its efficiency improvements at the cost of small overhead. We embedded Spray in a real-time decentralized editor running in browsers and ran experiments involving up to 600 communicating web browsers. The results demonstrate that Spray significantly reduces the network traffic according to the number of participants and saves bandwidth.

Genet: A Quickly Scalable Fat-Tree Overlay for Personal Volunteer Computing using WebRTC

“…The most promising one seems to bootstrap other overlay networks built with WebRTC. It could, for example, implement a peer sampling protocol, such as Spray [36], and the initial bootstrap could be made fast by having new nodes join multiple nodes in the tree, forming a mesh that could then progressively converge to an efficient topology. The quick scaling ability of the design we have presented is therefore complementary to potential refinements based on existing overlay designs.…”

Breaking the Scalability Barrier of Causal Broadcast for Large and Dynamic Systems

“…We expect the delay to increase as the latency increase. Description: We build an overlay network with a topology close to random graphs using Spray [24]. The overlay networks comprises 1k, and 10k processes.…”

Foiling Sybils with HAPS in Permissionless Systems: An Address-based Peer Sampling Service

“…This idea has since been extended to the implementation of efficient Byzantine-resilient primitives in permission-less systems [6]. These promising techniques assume, however, a perfect Random Peer Sampling (RPS) service [7], [8], that is immune to Sybil attacks.…”